Full-Page Optical Music Recognition of Handwritten Monophonic Scores
Authors: Adrian Rosello, Antonio Ríos-Vila, David Rizo, Jorge Calvo-Zaragoza
Organizations: Pattern Recognition and Artificial Intelligence group, University of Alicante, Alicante, Spain · Instituto Superior de Enseñanzas Artísticas de la Comunidad Valenciana, Alicante, Spain
Full-page end-to-end Optical Music Recognition seeks to transcribe entire music pages directly into symbolic notation, avoiding the limitations of traditional pipelines that rely on accurate staff segmentation. Recent Transformer-based architectures have achieved strong performance on typeset scores, relying on large-scale synthetic data for pretraining. However, their applicability to handwritten music remains largely unexplored. In this work, we study full-page transcription on handwritten monophonic collections and analyze the impact of synthetic pretraining in this setting. To investigate which factors are most relevant during pretraining, we introduce a generator capable of producing visually coherent full-page scores in both typeset and handwritten styles. Experiments on three real handwritten datasets provide a comparative evaluation of several full-page pipelines and different synthetic pretraining strategies. The results suggest that the benefits of synthetic pretraining are primarily associated with learning structural layout conventions rather than with visual similarity to the target handwriting.
Figures & tables
Figure 1: 10-fold error rates after directly adjusting the Unfolding [ 10 ] and SMT [ 15 ] models to the target handwritten corpora.
Figure 2: Diagram of the two-step pipeline used.
Figure 3: Diagrams of the holistic pipelines used.
Figure 4: Representative examples of the generated music in both typeset (a) and handwritten (b) style.
Name
FMT-C
FMT-M
Jazzmus
Vocabulary size
60
75
96
Number of pages
86
205
163
Table 1: Statistics of the real full-page datasets.
Figure 5: Comparison of the SMT, Unfolding, and CRNNf scenarios in terms of SER across 10 folds for the FMT-C , FMT-M , and Jazzmus datasets.
FMT-C
FMT-M
Jazzmus
CRNNf
45.7±27.8
81.6±25.4
63.1±2.8
Unfolding
40.4±14.9
68.3±18.0
61.0±3.5
SMTpr
31.6±9.9
32.8±8.7
26.9±3.4
SMThw
38.4±10.4
32.9±5.2
30.6±3.1
Table 2: 10-fold cross-validated SER (%) on FMT-C , FMT-M , and Jazzmus . Values report mean and standard deviation across folds. Best results per dataset are in bold.
Figure 6: Qualitative result of the model SMTpr performance on a test split of the 10k-fold of Jazzmus . Red denotes false positives, while blue represents false negatives.